Michael Glanzberg

The Taming of the True continues the project Neil Tennant began in Anti-realism and Logic of investigating and defending anti-realism. Tennant’s earlier book anticipated a second volume, in which issues related to empirical discourse would be addressed in greater detail. The Taming of the True provides this sequel. It also attempts a ground-clearing project, by addressing challenges to some of the presuppositions and implications of Tennant’s anti-realist position. Finally, it takes an opportunity to revisit some of the issues examined in the earlier book, in light of the evolving debate over realism and anti-realism.

This paper argues for the thesis that, roughly put, it is impossible to talk about absolutely everything. To put the thesis more precisely, there is a particular sense in which, as a matter of semantics, quantifiers always range over domains that are in principle extensible, and so cannot count as really being ‘absolutely everything’. The paper presents an argument for this thesis, and considers some important objections to the argument and to the formulation of the thesis. The paper also offers an assessment of just how implausible the thesis really is. It argues that the intuitions against the thesis come down to a few special cases, which can be given special treatment. Finally, the paper considers some metaphysical ideas that might surround the thesis. Particularly, it might be maintained that an important variety of realism is incompatible with the thesis. The paper argues that this is not the case.

Current theories of context see context as composed of information that is localizable to individual utterances. Current theories of discourse grant that discourses have important global properties that are not so localizable. In this paper, I argue that context, even narrowly construed as whatever combines with a sentence to determine truth conditions, must have a discourse–global component. I identify a context–dependence phenomenon related to the linguistic concepts of topic and focus, isolate the pertinent feature of context, and show that this feature must be discourse–global in nature. I thus argue that context is as complicated as an entire discourse.

Quantification is haunted by the specter of paradoxes. Since Russell, it has been a persistent idea that the paradoxes show what might have appeared to be absolutely unrestricted quantification to be somehow restricted. In the contemporary literature, this theme is taken up by Dummett (1973, 1993) and Parsons (1974a,b). Parsons, in particular, argues that both the Liar and Russell’s paradoxes are to be resolved by construing apparently absolutely unrestricted quantifiers as appropriately restricted.

This paper reviews the claim that certain predicates, including what are called predicates of personal taste, have a sometimes-hidden element for a judge or experiencer. This claim was advanced in my own earlier work, as well as a number of other papers. My main goal here is to review some of the arguments in favor of this claim, and along the way, to present some of my earlier unpublished work on the matter. In much of the earlier literature, this claim was part of a debate between relativists, contextualists, and others about the semantics of ‘subjective’ or ‘perspectival’ predicates. I shall argue here that these issues are independent. Whether we opt for experiencer or judge parameters is independent of whether we prefer relativist semantics to any other kind.

Quantified terms are terms of generality. They are also provide some of our prime examples of the phenomenon of scope. The distinction between singular and general terms, as well as the ways that general terms enter into scope relations, are certainly fundamental to our understanding of language. Yet when we turn to natural language, we encounter a huge and apparently messy collection of general terms; not just every and some, but most, few, between five and ten, and many others. Natural-language sentences also display a complex range of scope phenomena; unlike first-order logic, which clearly and simply demarcates scope in its notation.

This paper revisits a challenge for contextualist approaches to paradoxes such as the Liar paradox and Russell’s paradox. Contextualists argue that these paradoxes are to be resolved by appeal to context dependence. This can offer some nice and effective ways to avoid paradox. But there is a problem. Context dependence is, at least to begin with, a phenomenon in natural language. Is there really such context dependence as the solutions to paradoxes require, and is it really just a familiar linguistic phenomenon at work? Not so clearly. In earlier work, I argued that the required form of context dependence does not look like our most familiar instances of context dependence in natural language. I called this extraordinary context dependence. In this paper, I shall explore, somewhat tentatively, a way that we can see the context dependence needed to address paradoxes as not so extraordinary. Doing so will also allow us to connect thinking about the context dependence of quantifier domains with some interesting ideas about the distinctive semantic properties of certain quantifiers.

This chapter reviews Davidson's main work on truth. It focuses on the connections between truth, meaning, and interpretation that form the core of Davidson's views, and on the relations of his views to traditional theories of truth. It highlights several key ideas that comprise Davidson's approach to truth: Tarski's work on truth is fundamental to understanding the concept, as is the relation of truth to meaning, and we fail to understand that connection adequately unless we take into account the constraints of radical interpretation. Furthermore, in understanding those connections, we see that truth is not a redundant or trivial concept, but equally, it is not a locus for metaphysical commitments.

Semanticists will often casually remark that Noam Chomsky rejects semantics. Chomsky has frequently noted how poorly understood some aspects of semantics are, and has shown little inclination to grant the status of reasonably well‐developed science to many parts of semantics. One specific reason Chomsky has often voiced skepticism about semantics is that he saw the wrong kinds of appeals to semantics in the wrong places. The arguments for the autonomy of syntax in Chomsky's early writing have been described as part of an extended argument for the “existence of syntax”. Semantics for Chomsky must be part of I‐language: the part that provides useful information from the Language Faculty to the Conceptual‐Intentional system. There are many traditions in semantics, and many current research programs in sematics are embedded in firmly anti‐Chomskyan views of language in general. Chomsky's skepticism about truth‐conditional semantics has been, and remains strong.

One of the great successes in the study of language has been the application of formal methods, including those of formal logic. Even so, this chapter argues against one way of accounting for this success, by arguing that the study of natural language semantics and of logical consequence relations are not the same. There is indeed a lot we can glean about logic from looking at our languages, and at our inferential practices, but the semantic properties of natural languages do not determine genuine logical consequence relations. We can get from natural language semantics to logical consequence, but only by a significant process of identification of logical constants, abstraction, and idealization. The chapter also discusses different approaches to the nature of logical consequence, and examines which allow logic and natural language to come closer together.

Gary Ostertag: Quine and Russell: The chapter provides a selective overview of themes common to Russell and Quine, focusing on Russell's theory of descriptions and the notion of contextual definition. It begins by discussing Russell and Quine on modality, along the way highlighting the following topics: how C.I. Lewis's metalinguistic understanding of the modal operators shaped the subsequent debate about modality – in particular, how it rendered the very idea of de re modality unintelligible; how Quine's inattention to matters of scope led him to conclude, somewhat misleadingly, that he had derived a modal paradox; how the failure of the reductionist program provided the basis for a challenge to a cornerstone of the understanding of modality – the idea of analytic truth. I then turn to Russell and Quine on the semantics of attitude reports, where the issue of scope takes center stage again. A final section discusses Russell's influence in the shaping of Quine's ontological criterion.

One of the difficulties surrounding the ancient Paradox of the Liar is that it is unclear what philosophical problem it poses. In this thesis, I offer a diagnosis of the problem, and defend a solution to it. In Chapter I, I argue that the problem posed is one about the relation between a sentence uttered in a context and the proposition expressed. Chapter II supports this conclusion, by arguing against most notions of truth-value gap, and by elaborating on the notion of failing to express a proposition. ;To develop a solution to the problem, I argue in Chapter III that the notion of context must be expanded to take into account structural features of the discourse in which an utterance is made. So understood, a context determines a domain from which the truth conditions available for the construction of propositions expressible in the context must be drawn. ;This theory is applied to the Liar Paradox in Chapter IV. Under my enhanced notion of context, we can see certain inferences as changing a context just by introducing novel words into a conversation. In the Liar Paradox case, the term for the semantic relation of expressing a proposition is introduced. Because such relations have a highly complex local character, this context-change induces a substantial change in the domain of available truth conditions. The result is, I show, that there are effectively more propositions that can be expressed in the new context than there were in the old. Among the new propositions that can be identified in the new context is the one expressed by the Liar Sentence. ;One of the most striking features of a hierarchical account of this kind is the result that any quantification, whether syntactically unrestricted or not, is interpreted as being over some restricted domain. I go on to argue in Chapter V that in fact there are independent reasons for thinking that quantification must behave in this way. ;Finally, I elaborate some of the more general consequences of the theory I develop. First, I claim that the notion of expression I propose undermines any sharp and substantial distinction between semantics and pragmatics. Second, in Chapter VI, I argue that the theory has some consequences for contemporary debates over realism and truth